Flow Lab

Syllabus / Module 15: Applications: Biology, Chemistry & Society

Lesson 15.6

Neuron Models

Hodgkin–Huxley, FitzHugh–Nagumo, excitability, action potentials, and the cable equation.

Written byClaude Sonnet

🌊
Why this matters for simulation: Excitable media waves look strikingly like combustion fronts.

By the end of this lesson you can…

  • Derive the leaky integrate-and-fire firing rate and its threshold current.
  • Explain all-or-none spikes, refractoriness and the jump to repetitive firing from Hodgkin-Huxley numbers.
  • Draw the FitzHugh-Nagumo nullclines and find where the rest point loses stability in the injected current.
  • Find the firing threshold numerically and say why a kick must exceed it.
  • Measure the speed of a FitzHugh-Nagumo pulse with diffusion and explain why it is not the Fisher-KPP speed.
  • Choose a step size for the stiff spike upstroke and say how Euler and RK4 fail.
On this page
·
  1. The membrane as a circuit
  2. Leaky integrate-and-fire
  3. Hodgkin-Huxley
  4. Reduction: FitzHugh-Nagumo
  5. The cable equation and a travelling pulse
  6. Numerics: a mildly stiff upstroke
  7. What the models leave out
  8. Where this goes next
  9. The exercise: rate, rest point, Hopf current and threshold
  10. Concept checks
  11. Problems
  12. Key Takeaways

15.5 Reaction–Diffusion & Turing Patterns ended on a promise: a fast voltage variable and a slow recovery variable, diffusing in space, turn a front into a travelling pulse. This lesson builds the neuron versions of that pair. We start with the cheapest spiking model (a leaky capacitor with a reset), then the 1952 Hodgkin-Huxley equations, then the two-variable cartoon that lets us draw everything in a phase plane. The Hopf bifurcation we met in 15.4 Chemical Kinetics & Stiff Systems returns as the switch from silence to repetitive firing, and the cable equation gives the pulse a speed that we can measure.

▶ Video A leaky capacitor and its firing rate, the Hodgkin-Huxley spike and its gates, the FitzHugh-Nagumo phase plane with nullclines, a pulse travelling down a cable, and why the upstroke is stiffcaptions only
Transcript
  1. The last lesson promised a pulse. A membrane is a capacitor with ion channels, C dV dt equals I minus I ion, and a single linear leak only relaxes, it never spikes. A fast voltage and a slow recovery, diffusing in space, should turn a front into a travelling pulse. So we need reaction terms: the cheapest spiking rule, the 1952 Hodgkin and Huxley equations, then a two variable cartoon we can draw in a plane. Each is a choice of I ion.
  2. Leaky integrate and fire keeps the leak and adds a rule: at threshold, spike and reset. The steady voltage is E L plus R I, so below the threshold current of 1.5 nanoamps it never fires. Above it, the exact solution reaches threshold at tau log of V infinity minus reset over V infinity minus threshold. At 2 nanoamps that is 13.86 milliseconds, 72.13 hertz, or 63.04 with a 2 millisecond refractory time. The rate rises continuously from zero, so arbitrarily slow firing is possible. The next model will not do this.
  3. Hodgkin and Huxley fitted three gates to the squid axon: sodium opens fast with m, inactivates slowly with h, and potassium opens slowly with n. Rest sits at minus 64.996 millivolts. A kick of 6.5 dies away, peaking at minus 56.0; a kick of 7 fires a spike to 37.1. The threshold is 6.505: all or none. A 10 millivolt kick peaks at 39.4 after 1.78 milliseconds, then undershoots to minus 76.2, near E K. For about 5 milliseconds a second kick cannot fire: it needs 46.5 at 6, 27 at 8, 17 at 10, 7 at 15, and 6 at 20.
  4. Now hold a current on. The rheobase is 2.240: one spike, then silence. Repetitive firing needs more. At I equal 6.2 the rate is zero; at 6.3 it is 52.37 hertz, and it cannot be slower. That is the class two onset, unlike integrate and fire. Beyond, 7 gives 58.33, 10 gives 68.32, and 20 gives 86.47. The Jacobian's leading pair crosses the imaginary axis at 9.7754, a Hopf bifurcation. But a spike train exists below it, down to 6.260, so rest and firing coexist and the Hopf looks subcritical. That last part is my inference.
  5. FitzHugh and Nagumo built a caricature: a fast cubic v and a slow linear recovery w. The v nullcline is a cubic, the w nullcline a line, and they cross at the rest point, minus 1.1994, minus 0.6243. Trace minus 0.5026 and determinant 0.1081 make a stable spiral. A kick of 0.5 peaks at minus 0.670 and dies; 0.6 peaks at 1.714 and fires. Bisection puts the threshold at 0.5555, a little right of the unstable middle root, minus 0.7863. Past it, v runs to the right branch, then w rises and returns.
  6. Linearise: the trace is 1 minus v star squared minus epsilon b, zero at v star squared equal 0.936. That gives two Hopf currents, 0.33128 and 1.41872, and between them the rest point is unstable and the system fires periodically. At I equal 0.5, 1.0 and 1.4 the amplitude is 3.82, 3.84 and 3.78, with periods 39.47, 36.70 and 45.61. A small Hopf oscillation would have period 22.8, but at 0.34 the sim shows 46.79: the full cycle appears almost at once. It even survives down to 0.3242.
  7. Current also flows sideways. The cable equation makes voltage diffuse with D equal r over 2 R i C, while the channels react: exactly last lesson's form, with only v diffusing. Start FitzHugh and Nagumo with v equal 2 on the left. A pulse forms and runs right, shape unchanged, at speed 0.803. Quadrupling D gives 1.617, a ratio of 2.01: speed goes as root D. It is not the Fisher speed: ahead of the pulse the rest state is stable, so the front is pushed. With recovery frozen the speed is 0.9630, and faster recovery slows the pulse.
  8. Run Hodgkin and Huxley at I equal 10. RK4 with step 0.001 gives the first spike at 1.9012 milliseconds, and seven spikes in 100. Euler is first order: its first spike comes at 1.9029, 1.9179 and 1.9809 for steps 0.001, 0.01 and 0.05, late by about 1.7 steps, while RK4 stays at 1.9012. At 0.1 both diverge. The cause is the upstroke, where the Jacobian's largest eigenvalue is 36.4 per millisecond at 2.28. That caps Euler near 0.055 and RK4 near 0.077; bisection found 0.0728 and 0.0939. This is mild stiffness.
  9. Hodgkin and Huxley describes one channel set in a squid axon at one temperature. Real neurons have dozens of channel types, dendrites, synapses and noise, so the threshold we found to nine digits is, in a cell, a probability. Every number here belongs to an equation set, not a neuron. A cable pulse moves by diffusion plus reaction, at a speed that goes as the square root of D. Next, traffic drops diffusion and asks about conservation: cars satisfy rho t plus the derivative of rho times v of rho equals 0. Waves steepen into shocks. What plays the role of the threshold?

The membrane as a circuit

A patch of membrane is a capacitor CC in parallel with ion channels. Charge conservation gives

C dVdt=I−Iion(V,gates),C\,\frac{dV}{dt} = I - I_{\mathrm{ion}}(V, \text{gates}),

with II the injected current. Everything below is a choice of IionI_{\mathrm{ion}}. A single linear leak gives the RC circuit of 14.2 Electrical Circuits: first order, relaxing to a steady voltage, never spiking. Spikes need nonlinear channels, or a rule added by hand.

Leaky integrate-and-fire

The cheapest model keeps the leak and adds a rule: when VV reaches VthV_{th}, record a spike and reset to VresetV_{reset}. With R=1/gLR = 1/g_L,

τ V˙=−(V−EL)+RI,V∞=EL+RI.\tau\,\dot V = -(V - E_L) + RI, \qquad V_\infty = E_L + RI.

If V∞≤VthV_\infty \le V_{th} the voltage settles below threshold forever. So the threshold current is Ith=(Vth−EL)/RI_{th} = (V_{th} - E_L)/R. Above it, the exact solution from VresetV_{reset} is V(t)=V∞+(Vreset−V∞)e−t/τV(t) = V_\infty + (V_{reset} - V_\infty)e^{-t/\tau}, which reaches VthV_{th} at

T=τln⁡V∞−VresetV∞−Vth,f=1tref+T,T = \tau \ln\frac{V_\infty - V_{reset}}{V_\infty - V_{th}}, \qquad f = \frac{1}{t_{ref} + T},

with an optional refractory time treft_{ref} during which the voltage is held at reset. Take τ=10\tau = 10 ms, EL=Vreset=−65E_L = V_{reset} = -65 mV, Vth=−50V_{th} = -50 mV, R=10R = 10 MΩ\Omega. Then Ith=1.5I_{th} = 1.5 nA. At I=2I = 2 nA, V∞=−45V_\infty = -45 mV and T=10ln⁡4=13.86T = 10\ln 4 = 13.86 ms, so f=72.13f = 72.13 Hz; with tref=2t_{ref} = 2 ms, f=63.04f = 63.04 Hz.

II (nA)treft_{ref} (ms)formula (Hz)simnumpy Euler
1.61.60036.0736.0736.0736.07
220072.13572.13572.13472.13472.13472.134
222263.04063.04063.03663.03663.03863.038

The rate rises continuously from zero at IthI_{th} (the logarithm diverges as V∞→VthV_\infty \to V_{th}), so arbitrarily slow firing is possible; the next model will not do this. The model has no spike shape, and its reset is a discontinuity: a hybrid system, not a smooth ODE.

◆ Interactive Leaky integrate-and-fire: raise the current past the threshold and watch the rate curve

Loading simulation…

Hodgkin-Huxley

Hodgkin and Huxley (1952) fitted ODEs to voltage-clamp currents from the squid giant axon:

CV˙=I−gNa m3h (V−ENa)−gK n4 (V−EK)−gL(V−EL),C\dot V = I - g_{Na}\,m^3h\,(V - E_{Na}) - g_K\,n^4\,(V - E_K) - g_L(V - E_L),

and each gating variable x∈{m,h,n}x \in \{m, h, n\}, a probability in [0,1][0, 1], obeys x˙=αx(V)(1−x)−βx(V) x\dot x = \alpha_x(V)(1 - x) - \beta_x(V)\,x. Sodium opens fast (mm) then inactivates slowly (hh); potassium opens slowly (nn). Positive feedback from mm makes the upstroke; hh and nn are the delayed negative feedback that ends it.

I use the modern convention (rest near −65-65 mV): C=1C = 1 μ\muF/cm2^2, gNa=120g_{Na} = 120, gK=36g_K = 36, gL=0.3g_L = 0.3 mS/cm2^2, ENa=50E_{Na} = 50, EK=−77E_K = -77, EL=−54.387E_L = -54.387 mV, with the standard rate functions, e.g. αm=0.1(V+40)/(1−e−(V+40)/10)\alpha_m = 0.1(V + 40)/(1 - e^{-(V+40)/10}) and βm=4e−(V+65)/18\beta_m = 4e^{-(V+65)/18}. All of these constants are from memory. What I can check is what they produce: every number below comes from the sim (RK4) and an independent scipy or numpy solve.

Rest. Solving Iion(V)=0I_{ion}(V) = 0 at steady-state gates gives V=−64.996V = -64.996 mV with (m,h,n)=(0.0530,0.5960,0.3177)(m, h, n) = (0.0530, 0.5960, 0.3177). Landing on −65-65 suggests the constants are right.

All or none. A voltage kick δ\delta from rest dies away for small δ\delta and fires a full spike above δ∗=6.505\delta^* = 6.505 mV: at δ=6.5\delta = 6.5 the voltage never exceeds −56.0-56.0 mV, at δ=7\delta = 7 it reaches 37.137.1 mV. A 1010 mV kick peaks at +39.4+39.4 mV after 1.781.78 ms, then undershoots to −76.2-76.2 mV at 4.624.62 ms, near EKE_K because potassium is still open (scipy: 39.42839.428 and −76.17-76.17).

Refractoriness. After a spike hh is small and nn large. I applied a second kick at a delay after the upward zero crossing and searched, in 0.50.5 mV steps up to 6060, for the smallest one that fires again:

delay (ms)44556688101015152020
smallest kick (mV)nonenone46.546.5272717177766

Sim and scipy agree on every entry: about 55 ms of deafness, then gradual recovery as hh climbs back to 0.60.6.

Current steps. A sustained current II (µA/cm2^2) from rest gives a spike once I≥2.240I \ge 2.240 (the rheobase over 200200 ms; sim and scipy 2.24032.2403). At I=3I = 3 and 55 a 400400 ms step gave a single spike, then silence. Repetitive firing needs more:

II6.26.26.36.377101020205050
rate (Hz)0052.3752.3758.3358.3368.3268.3286.4786.47117.04117.04

Between 6.26.2 and 6.36.3 the rate jumps from 00 to about 5252 Hz and cannot be slower. Neurons like this are called class II (term from memory); integrate-and-fire is class I.

Which bifurcation? The 4×44\times4 Jacobian at the equilibrium has a leading eigenvalue pair that crosses the imaginary axis at I=9.7754I = 9.7754 (sim and scipy): a Hopf bifurcation, as in 3.9 Bifurcations and 7.9 Limit Cycles & Hopf Bifurcations. But firing exists below that. Start on the cycle at I=10I = 10 and lower the current: it survives down to I=6.260I = 6.260 (a saddle-node of cycles) and then vanishes. For 6.26<I<9.786.26 \lt I \lt 9.78 stable rest and a stable spike train coexist, so the Hopf looks subcritical; I did not compute the unstable cycle between them, so that is an inference.

◆ Interactive Hodgkin-Huxley: step the current or kick the voltage, and watch V and the gates m, h, n

Loading simulation…

Reduction: FitzHugh-Nagumo

FitzHugh (1961) and Nagumo and colleagues (1962; dates from memory) built a caricature: a fast cubic voltage-like vv and a slow linear recovery ww,

v˙=v−v33−w+I,w˙=ε (v+a−b w),a=0.7, b=0.8, ε=0.08\dot v = v - \frac{v^3}{3} - w + I, \qquad \dot w = \varepsilon\,(v + a - b\,w), \quad a = 0.7,\ b = 0.8,\ \varepsilon = 0.08

(the textbook constants, from memory). The vv-nullcline is the cubic w=v−v3/3+Iw = v - v^3/3 + I, the ww-nullcline the line w=(v+a)/bw = (v + a)/b, as in 7.3 Phase Portraits & the Trace–Determinant Plane. They cross at the rest point; for these a,ba, b the root is unique.

Excitability. At I=0I = 0 the rest point is (v,w)=(−1.1994,−0.6243)(v, w) = (-1.1994, -0.6243) with trace −0.5026-0.5026, determinant 0.10810.1081 and eigenvalues −0.2513±0.2119 i-0.2513 \pm 0.2119\,i (1.7 Complex Eigenvalues & Rotation): a stable spiral. A kick δv=0.5\delta v = 0.5 peaks at v=−0.670v = -0.670; δv=0.6\delta v = 0.6 peaks at 1.7141.714. Bisection puts the threshold at δv∗=0.5555\delta v^* = 0.5555, i.e. v=−0.6439v = -0.6439 (sim and scipy to 10−910^{-9}). At the rest height w=−0.6243w = -0.6243 the cubic has roots −1.1994-1.1994 (rest), −0.7863-0.7863 (the unstable middle branch) and 1.98571.9857. Past the middle branch the fast variable runs to the right branch (the spike), then ww slowly rises and the state returns along the left branch. The threshold lies a little right of −0.7863-0.7863; I did not derive the gap.

Loss of stability. Linearising (7.7 Nonlinear Systems & Linearization), the trace is 1−v∗2−εb1 - v^{*2} - \varepsilon b, which vanishes at v∗2=1−εb=0.936v^{*2} = 1 - \varepsilon b = 0.936. Putting v∗=∓0.9675v^* = \mp 0.9675 into I=−v∗+v∗3/3+(v∗+a)/bI = -v^* + v^{*3}/3 + (v^* + a)/b gives the two Hopf currents

IH1=0.33128,IH2=1.41872,I_{H1} = 0.33128, \qquad I_{H2} = 1.41872,

confirmed by bisection and scipy to 10−1210^{-12}. Between them the rest point is unstable and the system fires periodically. At I=0.5,1.0,1.4I = 0.5, 1.0, 1.4 the cycle has amplitude 3.82,3.84,3.783.82, 3.84, 3.78 and period 39.47,36.70,45.6139.47, 36.70, 45.61 (sim and scipy agree to 10−610^{-6}).

How abrupt is the onset? A small Hopf oscillation just past IH1I_{H1} would have ω=det⁡J=0.2755\omega = \sqrt{\det J} = 0.2755 and period 22.822.8. The sim shows at I=0.34I = 0.34 amplitude 3.773.77 and period 46.7946.79: the full cycle appears within about 0.010.01 of II. Started on the large cycle and lowered, it survives down to I=0.3242I = 0.3242 (sim 0.324180.32418, scipy 0.324190.32419), below the Hopf, so there is a tiny bistable window. Such abrupt growth at small ε\varepsilon is called a canard explosion (term from memory; not analysed).

◆ Interactive FitzHugh-Nagumo phase plane: kick from rest, click anywhere to start a trajectory, slide I past 0.33

Loading simulation…

The cable equation and a travelling pulse

Current also flows sideways along an axon. Balancing charge on a thin cylinder of radius rr and axial resistivity RiR_i gives the cable equation (derivation from memory)

C ∂tV=r2Ri ∂x2V+I−Iion,C\,\partial_t V = \frac{r}{2R_i}\,\partial_x^2 V + I - I_{\mathrm{ion}},

so voltage diffuses with D=r/(2RiC)D = r/(2R_iC) while the channels react: exactly ∂tu=D∇2u+f(u)\partial_t\mathbf u = D\nabla^2\mathbf u + \mathbf f(\mathbf u) of 15.5 Reaction–Diffusion & Turing Patterns, with only vv diffusing. Take FHN at I=0I = 0, D=1D = 1 on [0,200][0, 200] with no-flux ends, Δx=0.5\Delta x = 0.5, explicit step 0.020.02, and start with v=2v = 2 on x<5x \lt 5.

A pulse forms and runs right, shape unchanged. Tracking where vv falls through zero between t=40t = 40 and 8080, the speed is 0.8030.803 (scipy on the same grid: 0.8040.804). With Δx=0.25\Delta x = 0.25 and the time step cut to 0.0050.005 it is 0.8100.810 (scipy 0.8100.810), so about 0.80.8 to one percent. Quadrupling DD gives 1.6171.617 (scipy 1.6201.620), a ratio of 2.012.01: speed goes as D\sqrt D, as rescaling xx by D\sqrt D demands.

Fisher or not? The Fisher speed 2Dr2\sqrt{Dr} came from linearising where the state is unstable. Ahead of this pulse the rest state is stable: ∂f/∂v=1−vrest2=−0.4386\partial f/\partial v = 1 - v_{rest}^2 = -0.4386. A linear tail can only decay, so there is no c(λ)c(\lambda) to minimise. The front is pushed: the nonlinearity as a whole sets the speed. To test that, freeze recovery (ε=0\varepsilon = 0, ww at rest), so vt=Dvxx−13(v−v1)(v−v2)(v−v3)v_t = Dv_{xx} - \tfrac13(v - v_1)(v - v_2)(v - v_3) with roots −1.1994,−0.7863,1.9857-1.1994, -0.7863, 1.9857. The standard cubic front speed (formula from memory) is D/6 (v1+v3−2v2)=0.9630\sqrt{D/6}\,(v_1 + v_3 - 2v_2) = 0.9630 at D=1D = 1. Measured: 0.9520.952 (sim) and 0.9560.956 (scipy) at Δx=0.5\Delta x = 0.5; 0.9600.960 and 0.9610.961 at 0.250.25. It converges to the formula.

With recovery on, the pulse is slower, and the faster the recovery the slower it goes:

ε\varepsilon0.020.020.040.040.080.080.100.10frozen
sim0.92150.92150.88730.88730.80300.80300.74100.74100.96030.9603
scipy0.92510.92510.89030.89030.80430.80430.74100.74100.96140.9614

(Frozen column at Δx=0.25\Delta x = 0.25 with time step 0.0050.005.) With this stimulus no pulse forms for ε≥0.12\varepsilon \ge 0.12; at 0.110.11 sim and scipy disagree because it is marginal, and I did not try a bigger stimulus.

◆ Interactive A travelling FitzHugh-Nagumo pulse: change D and eps and read the measured speed

Loading simulation…

Numerics: a mildly stiff upstroke

Run HH at I=10I = 10 for 100100 ms. RK4 with Δt=0.001\Delta t = 0.001 ms gives the first spike at 1.90121.9012 ms, seven spikes and peak 40.2640.26 mV. Forward Euler (4.2 The Euler Family: Forward, Backward, Heun & Midpoint):

Δt\Delta t (ms)0.0010.0010.010.010.050.050.10.1
Euler first spike (ms)1.90291.90291.91791.91791.98091.9809diverges
RK4 first spike (ms)1.90121.90121.90121.90121.90121.9012diverges

(My numpy loops agree with the sim to these digits.) Euler is first order: the first spike is late by about 1.7 Δt1.7\,\Delta t. At 0.10.1 ms Euler overshoots to 215215 mV, past ENaE_{Na}, then diverges. The cause is the Jacobian along the spike: its largest ∣Re λ∣|\mathrm{Re}\,\lambda| is 36.436.4 per ms at t=2.28t = 2.28 ms (sim and scipy 36.3636.36), on the upstroke. Linear stability needs h<2/36.4=0.055h \lt 2/36.4 = 0.055 ms for Euler and 2.785/36.4=0.0772.785/36.4 = 0.077 ms for RK4 (4.5 Stability Regions & Stiffness, 4.3 Runge–Kutta Methods & Butcher Tableaux). Bisecting on “no blow-up and 0<Vmax<1000 \lt V_{max} \lt 100” found limits of 0.07280.0728 (Euler) and 0.09390.0939 (RK4), a little above those estimates; I did not isolate why, perhaps because the large eigenvalue is visited only briefly. This is mild stiffness: the fast scale caps hh at a few hundredths of a millisecond while a spike lasts a few milliseconds, and RK4 at 0.010.01 is accurate. A stiff solver pays off for long runs or cables with thousands of nodes.

What the models leave out

Hodgkin-Huxley describes one channel set in a squid axon at one temperature. Real neurons have dozens of channel types (which can turn class II into class I), dendrites that are cables with their own channels, synapses, and noise: a few thousand channels flip randomly, so the threshold we found to nine digits is, in a real cell, a probability. Every number above is a property of an equation set, not of a neuron.

Where this goes next

A cable pulse moves by diffusion plus reaction: information travels because a neighbour is pulled across threshold, and the speed scales as D\sqrt D. The next lesson, 15.7 Traffic Flow, drops diffusion and asks about conservation: cars on a road satisfy ∂tρ+∂x(ρ v(ρ))=0\partial_t\rho + \partial_x(\rho\,v(\rho)) = 0, a first-order PDE with a nonlinear flux. Waves there travel along characteristics (9.10 First-Order PDEs: Advection & Characteristics), can steepen into shocks (9.11 Nonlinear Waves: Burgers' Equation & Shocks), and need the finite-volume ideas of 10.10 The Finite Volume Method. A traffic jam is not a neuron. But it is a nonlinear wave that holds its shape and travels, and we can ask what plays the role of the threshold.

The exercise: rate, rest point, Hopf current and threshold

The playground integrates Hodgkin-Huxley with RK4 and with Euler. In the exercise you build the leaky integrate-and-fire rate, the FitzHugh-Nagumo rest point, the trace, the Hopf current by bisection, and a spike detector by RK4.

🐍 PythonHodgkin-Huxley: RK4 against forward Euler
⌘/Ctrl + Enter
🧪 Coding exerciseExercise: LIF rate, FitzHugh-Nagumo rest point, Hopf current and spike peak
⌘/Ctrl + Enter

Concept checks

✅ Concept check

A leaky integrate-and-fire neuron has $E_L = -70$ mV, $V_{th} = -55$ mV and $R = 30$ M$\Omega$. What is its threshold current, in nA?

✅ Concept check

In the lesson's Hodgkin-Huxley numbers, the firing rate goes from zero at $I = 6.2$ to $52$ Hz at $I = 6.3$. What does this tell you compared with integrate-and-fire?

✅ Concept check

For FitzHugh-Nagumo with $\varepsilon = 0.08$, $b = 0.8$, the trace of the Jacobian is $1 - v^{*2} - \varepsilon b$. At the Hopf points, what is $v^{*2}$?

✅ Concept checkselect all that apply

Select all true statements about the FitzHugh-Nagumo pulse speed in this lesson.

✅ Concept check

The largest $|\mathrm{Re}\,\lambda|$ along an HH spike is $36.4$ per ms. What step does the linear stability bound $2/|\lambda|$ for forward Euler allow, in ms to two decimals?

Problems

✏️ Pen & paperRate with a refractory period●○○

A leaky integrate-and-fire neuron has τ=10\tau = 10 ms, EL=Vreset=−65E_L = V_{reset} = -65 mV, Vth=−50V_{th} = -50 mV, R=10R = 10 MΩ\Omega and tref=2t_{ref} = 2 ms. Find its rate at I=4I = 4 nA and the largest rate it can reach.

Show worked solution

V∞=−25V_\infty = -25 mV, so T=10ln⁡4025=4.700T = 10\ln\frac{40}{25} = 4.700 ms and f=1000/(2+4.700)=149.25f = 1000/(2 + 4.700) = 149.25 Hz (sim and numpy agree). As I→∞I \to \infty, T→0T \to 0 and f→1/tref=500f \to 1/t_{ref} = 500 Hz.

✏️ Pen & paperThe Hopf current by hand●●○

For FitzHugh-Nagumo with a=0.7a = 0.7, b=0.8b = 0.8, ε=0.08\varepsilon = 0.08: (a) show the trace vanishes at v∗=±0.9675v^* = \pm 0.9675; (b) find the two currents; (c) find the small-oscillation period at the lower one.

Show worked solution

(a) 1−v∗2−εb=01 - v^{*2} - \varepsilon b = 0 gives v∗2=0.936v^{*2} = 0.936. (b) On the rest curve I=−v+v3/3+(v+a)/bI = -v + v^3/3 + (v + a)/b. For v=−0.96747v = -0.96747: 0.96747−0.30185−0.33434=0.33130.96747 - 0.30185 - 0.33434 = 0.3313. For v=+0.96747v = +0.96747: −0.96747+0.30185+2.08434=1.4187-0.96747 + 0.30185 + 2.08434 = 1.4187. (c) With 1−v∗2=εb1 - v^{*2} = \varepsilon b, det⁡=ε(1−εb2)=0.0759\det = \varepsilon(1 - \varepsilon b^2) = 0.0759, so ω=0.2755\omega = 0.2755 and the period is 2π/ω=22.82\pi/\omega = 22.8. The sim at I=0.34I = 0.34 shows 46.7946.79: the observed oscillation is not the small Hopf one.

✏️ Pen & paperA front from the middle root●●●

With recovery frozen, f(v)=v−v3/3−w=−13(v−v1)(v−v2)(v−v3)f(v) = v - v^3/3 - w = -\tfrac13(v - v_1)(v - v_2)(v - v_3) and the front speed is c=D/6 (v1+v3−2v2)c = \sqrt{D/6}\,(v_1 + v_3 - 2v_2) (from memory). (a) Show v1+v2+v3=0v_1 + v_2 + v_3 = 0, so c=−3D/2 v2c = -\sqrt{3D/2}\,v_2. (b) Evaluate it for middle root −0.7863-0.7863. (c) For which ww is the speed zero?

Show worked solution

(a) Expanding, −13(v−v1)(v−v2)(v−v3)=−13v3+13(v1+v2+v3)v2+…-\tfrac13(v - v_1)(v - v_2)(v - v_3) = -\tfrac13v^3 + \tfrac13(v_1 + v_2 + v_3)v^2 + \dots, and ff has no v2v^2 term, so the roots sum to 00. Then v1+v3=−v2v_1 + v_3 = -v_2 and c=D/6 (−3v2)=−3D/2 v2c = \sqrt{D/6}\,(-3v_2) = -\sqrt{3D/2}\,v_2. (b) 1.2247×0.7863=0.96301.2247 \times 0.7863 = 0.9630, matching the measured 0.9600.960 on the fine grid. (c) v2=0v_2 = 0 needs w=0w = 0, roots 0,±30, \pm\sqrt3: both wells equally deep, so neither state invades the other.

Key Takeaways

  • Leaky integrate-and-fire has the exact rate 1/(tref+τln⁡[(V∞−Vreset)/(V∞−Vth)])1/(t_{ref} + \tau\ln[(V_\infty - V_{reset})/(V_\infty - V_{th})]), threshold current 1.51.5 nA here, and a rate rising continuously from zero.
  • Hodgkin-Huxley (constants from memory) rests at −64.996-64.996 mV, fires all-or-none above a 6.5056.505 mV kick, peaks near +39.4+39.4 mV, undershoots to −76.2-76.2 mV and is deaf for about 55 ms.
  • Its rate jumps from 00 to about 5252 Hz near I=6.26I = 6.26 to 6.36.3; the rest point loses stability by a Hopf at 9.77549.7754, and between them rest and firing coexist.
  • FitzHugh-Nagumo reduces this to a plane: cubic nullcline, line, stable spiral at rest, kick threshold 0.55550.5555, Hopf currents 0.331280.33128 and 1.418721.41872 from v∗2=1−εbv^{*2} = 1 - \varepsilon b, with a full-amplitude cycle appearing at once.
  • With diffusion it gives a pulse of speed 0.800.80 that scales as D\sqrt D but is a pushed front, not the pulled Fisher front, because the rest state ahead is stable; as ε→0\varepsilon \to 0 the speed tends to the frozen front 3D/2 ∣v2∣=0.963\sqrt{3D/2}\,|v_2| = 0.963.
  • The upstroke is mildly stiff (largest decay rate 36.436.4 per ms): Euler fails above about 0.070.07 ms and is first order; RK4 fails above about 0.090.09 ms.
  • These are noise-free, single-channel-set models; the next lesson drops diffusion for a conservation law with shocks.
Finished this lesson?

Mark it complete to track your progress (saved in this browser). Reading every section does this for you.